Is it possible to use yolo for detecting pose of objects (not human body) from monocular rgbd camera?

[kappa95]

Hi community,

I have a silly and simple question. Is it possible to detect from video (from rgbd camera) an object (not human body or parts of human, only objects) and determine its pose (position and orientation in time) using yolo? I am not interested in real time application. if yes, are there any doc/guide explaining how to do that, and how to train the algorithm using videos/frames?

[UltralyticsAssistant]

👋 Hello @kappa95, thank you for reaching out and showing interest in Ultralytics 🚀!

While I can't address your question directly, I recommend checking out our Docs where you'll find a wealth of information that could be helpful for different use cases, including object detection with YOLO.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to assist us in debugging the issue.

For questions on custom training ❓, make sure to provide detailed information such as dataset image examples and training logs. You might also want to review our Tips for Best Training Results.

Engage with the Ultralytics community where it fits best for you! For quick chats, join us on Discord 🎧, or engage in detailed discussions on Discourse. Reddit fans can check out our Subreddit to share and gain insights with fellow users.

Upgrade

Make sure you are using the latest version of the ultralytics package with all requirements updated in a Python>=3.8 environment, with PyTorch>=1.8:

pip install -U ultralytics

Environments

YOLOv8 can run in these verified environments (dependencies such as CUDA/CUDNN, Python, and PyTorch are preinstalled):

• Notebooks:
• Google Cloud Deep Learning VM: GCP Quickstart Guide
• Amazon Deep Learning AMI: AWS Quickstart Guide
• Docker Image: Docker Quickstart Guide

Status

A green badge indicates that all Ultralytics CI tests are passing. These tests ensure correct functioning of all YOLOv8 Modes and Tasks across multiple platforms.

This is an automated response 🤖, but rest assured, an Ultralytics engineer will assist you shortly.

[glenn-jocher]

@kappa95 yOLOv8 is primarily designed for human pose estimation. For object pose estimation, you might need a specialized model or framework. Consider exploring other libraries or models tailored for object pose estimation tasks.

[kappa95]

Thank you Glenn, for the answer! Have you some suggestion?

[glenn-jocher]

@kappa95 you're welcome! For object pose estimation, consider exploring specialized models or frameworks tailored for that task.